Inhibition by active site directed covalent modification of human glyoxalase i

Article abstract of DOI:10.1016/j.bmc.2014.04.055

The glyoxalase pathway is responsible for conversion of cytotoxic methylglyoxal (MG) to d-lactate. MG toxicity arises from its ability to form advanced glycation end products (AGEs) on proteins, lipids and DNA. Studies have shown that inhibitors of glyoxa

Full text of DOI:10.1016/j.bmc.2014.04.055

Bioorganic & Medicinal Chemistry xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Inhibition by active site directed covalent modification of human
glyoxalase I
⇑
Ronald J. Holewinski , Donald J. Creighton
Department of Chemistry and Biochemistry, University of Maryland Baltimore County, 1000 Hilltop Circle, Baltimore, MD 21250, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The glyoxalase pathway is responsible for conversion of cytotoxic methylglyoxal (MG) to D-lactate. MG
Received 30 January 2014
Revised 18 April 2014
Accepted 28 April 2014
Available online xxxx
toxicity arises from its ability to form advanced glycation end products (AGEs) on proteins, lipids and
DNA. Studies have shown that inhibitors of glyoxalase I (GLO1), the first enzyme of this pathway, have
chemotherapeutic effects both in vitro and in vivo, presumably by increasing intracellular MG concentra-
tions leading to apoptosis and cell death. Here, we present the first molecular inhibitor, 4-bromoacetoxy-
1
-(S-glutathionyl)-acetoxy butane (4BAB), able to covalently bind to the free sulfhydryl group of Cys60 in
Keywords:
Glyoxalase
GLO1
Inhibitor
Advanced glycation end-products (AGEs)
Inactivator
Covalent modification
the hydrophobic binding pocket adjacent to the enzyme active site and partially inactivate the enzyme.
Our data suggests that partial inactivation of homodimeric GLO1 is due to the modification at only one of
the enzymatic active sites. Although this molecule may have limited use pharmacologically, it may serve
as an important template for the development of new GLO1 inhibitors that may combine this strategy
with ones already reported for high affinity GLO1 inhibitors, potentially improving potency and
specificity.
Ó 2014 Published by Elsevier Ltd.
1
. Introduction
conditions such as anxiety18–20 and depression.20 Distler et al.
A
identified MG as a GABA receptor agonist and reported that
The cellular glyoxalase system detoxifies cytosolic reactive
overexpression of GLO1 increases anxiety by reducing levels of
1
9
aldehydes that occur with metabolism. It is composed of the
MG and pretreatment with MG or inhibition of GLO1 reduced
1,2
21
enzymes glyoxalase I (GLO1, homodimer)
and glyoxalase II
GLO2, monomer) and is responsible for the conversion of the
methylglyoxal (MG) to -lactate via the intermediate S- -lactoyl-
pharmacologically-induced seizures in mice.
3
(
Additionally, increased GLO1 gene expression, protein expres-
sion, and activity have been reported in a variety of cancers, includ-
D
D
2
2–24
25
26
27–29
glutathione using catalytic amounts of glutathione (GSH,
ing breast,
pancreatic, melanoma, and prostate.
Vince
4
Scheme 1). MG is a small dicarbonyl that arises primarily through
and Daluge proposed that inhibitors of GLO1 could serve as anti-
tumor agents, by increasing concentrations of MG in tumor cells.30
S-substituted glutathione derivatives have been shown to be com-
the non-catalytic breakdown of triosephosphates (i.e., dihydroxy-
5,6
acetone phosphate and glyceraldehyde-3-phosphate) and from
7
the spontaneous breakdown of glucose. MG has the ability to form
petitive inhibitors of GLO1, with S-p-bromobenzylglutathione
8
,9
10
31
advanced glycation end-products (AGEs) on proteins,
lipids
i
being the most potent of these inhibitors (K = 0.08 lM). Thornal-
1
1
12
16,17
32
33
and DNA, resulting in detrimental effects in aging, as well in
ley et al. and Sakamoto et al. reported that S-p-bromobenzylg-
lutathione diesters had antitumor activity in human leukemia 60
cells and human lung cancer cells, respectively. Murthy et al.
diseases such as diabetes1
3–15
and Alzheimer’s disease,
where
GLO1 protein expression and activity have been shown to be
reduced. GLO1 and MG have also been linked to behavioral
i
reported the first transition state analogues of GLO1 with K values
3
4
ranging from 14 to 160 nM and Sharkey et al. reported that the
diethyl ester prodrug of S-(N-4-chlorophenyl-N-hydroxycarba-
moyl)glutathione (CHG) was able to inhibit tumor growth in
Abbreviations: CHG, S-(N-4-chlorophenyl-N-hydroxycarbamoyl)glutathione;
AGE, advanced glycation end product; MG, methylglyoxal; GSH, reduced glutathi-
one; GLO1, glyoxalase I.
3
5
mice. More recently, a variety of GLO1 inhibitors have been
⇑
Corresponding author at present address: Johns Hopkins University Bayview
synthesized, including the first bivalent-transition state analogues
Proteomics Center, Mason F. Lord Bldg, Center Tower, Room 602, Johns Hopkins
University, 5200 Eastern Ave, Baltimore, MD 21224, USA. Tel.: +1 4433880064.
E-mail address: ronholes7059@gmail.com (R.J. Holewinski).
Deceased.
36
i
of GLO1 (K 0.96–84 nM), in which two CHG molecules are
covalently linked together, enhancing the effective concentration
of inhibitor at the second active site. Additionally, More and Vince
http://dx.doi.org/10.1016/j.bmc.2014.04.055
0
968-0896/Ó 2014 Published by Elsevier Ltd.
Please cite this article in press as: Holewinski, R. J.; Creighton, D. J. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.055

2
R. J. Holewinski, D. J. Creighton / Bioorg. Med. Chem. xxx (2014) xxx–xxx
Scheme 1. Overview of the glyoxalase system.
have reported GLO1 inhibitors with K
i
values in the low micro/
addition of 4.44 g (1.92 mL, 0.022 mol) of bromoacetyl bromide,
3
7,38
nano-molar range
to -glutamyltranspeptidase, which is capable of cleaving the
-Glu-Cys bond of compounds such as S-p-benzoylglutathione and
that are metabolically stable and resistant
producing a yellow transparent solution. The solution was refluxed
for 20 h and the solvent was removed in vacuo using a Buchi
Rotavapor Re111 Evaporator yielding an orange tinted crystalline
solid. The crystals were re-dissolved in a minimum amount of
methanol and re-crystallized by the addition of water slowly until
c
c
3
9,40
the S-(N-aryl-N-hydroxycarbamoyl)glutathione derivatives.
To date, all of the GLO1 inhibitors above have been either
substrate or transition state analogues, and while some of these
molecules show strong potency, the potential to increase potency
and specificity should not be ignored. To directly address specific-
ity, we have taken a novel approach in designing an active-site
directed covalent inhibitor of GLO1. The covalent inhibitor was
designed based on the X-ray crystallographic structures of
the dimeric GLO1 with the substrate analogue S-p-bromobenzyl
a
white precipitate formed. The precipitate was vacuum
filtered and air dried overnight. The crystals were a flaky white
solid with an 82% percent yield and melting point of 60–62.5 °C.
1
The H NMR was consistent with the structure of BBE (d 1.75–
2 2 2
1.79 (t, –O–CH CH –); d 3.83 (s, Br–CH C(O)O–); d 4.19–4.23
(t, ꢀC(O)OCH –), Fig. S1). Four grams (12 mmol) of BBE was
2
dissolved in 150 mL of methanol with stirring. A glutathione
(GSH) solution was prepared by dissolving 0.8329 g (2.71 mmol)
2
glutathione and the transition state analogue S-(N-4-iodo-
1
phenyl-N-hydroxycarbamoyl)glutathione. These 3-dimensional
structures indicate that there is a hydrophobic binding pocket
composed of residues from both monomers (Met157, Leu160,
Phe162, Leu174, Met179, Met183 from one monomer; Cys60,
Phe62, Met65, Phe67, Leu69, Phe71, Ile88 from the other) adjacent
to the enzyme active site, with the free sulfhydryl group of C60
sticking into this hydrophobic pocket.
The goal of this study is to provide a proof of principle that
GLO1 could be inhibited covalently at Cys60 within the enzyme
active site and the aims of this study were threefold: (1) to synthe-
size a rudimentary GSH-analogue with a leaving group that could
potentially modify GLO1 covalently, (2) characterize the kinetic
parameters of this compound, and (3) determine the amino acid
modified in the active site. Here we report the inhibitor 4-bromo-
acetoxy-1-(S-glutathionyl)-acetoxy butane (4BAB, Fig. 1) that is
able to covalently modify Cys60 in the enzyme active site of GLO1.
2
2
. Experimental
.1. Materials and reagents
LC/MS solvents were obtained from Burdick and Jackson and
were of highest purity possible. Recombinant human GLO1 was
expressed and purified according to published methods and is
briefly described below.41 All other reagents were of the highest
grade possible.
2
.2. Synthesis of 4-bromoacetoxy-1-(S-glutathionyl)-acetoxy
butane (4BAB)
The synthetic route to 4BAB is outlined in Figure 1. First, the
starting di-ester bromoacetic acid 4-(2-bromo-acetoxy) butyl ester
(
1
BBE) was synthesized by dissolving 1.00 g (1.02 mL, 0.011 mol) of
,4-butanediol in 50 mL of methylene chloride followed by
Figure 1. Synthetic route to the GLO1 covalent inhibitor 4BAB.
Please cite this article in press as: Holewinski, R. J.; Creighton, D. J. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.055

R. J. Holewinski, D. J. Creighton / Bioorg. Med. Chem. xxx (2014) xxx–xxx
3
of GSH in 15 mL of degassed water and treating this solution with
.1 mL of 5 M NaOH to bring the pH of the solution to ꢁ9.5. The
2.4. Purification and esterification of GLO1 carboxymethylated
peptide C60-D-F-P-I-M-K66
1
GSH solution was added all at once to the BBE solution with
vigorous stirring. After two minutes, formic acid was added until
the pH of the solution was ꢁ3.5. The solvent was removed in vacuo
and a white solid remained. Water, 20 mL, was added to the solid
to dissolve 4BAB and any water-soluble impurities and 150 mL of
diethyl ether was added to dissolve the excess BBE. The two layers
from the filtrate were separated and the aqueous layer was saved
and washed once more with 150 mL of ether. The aqueous layer
was filtered through an aqueous syringe filter tip and 4BAB was
purified by on a Waters Delta 600 HPLC equipped with a Waters
The peak corresponding to the carboxymethyl modified peptide
60–66 (C60-K66CM, m/z 911.4, resulting from hydrolysis of the ester
bond of the 4BAB) from the GLO1:4BAB trypsin digest solution
(detailed below) was purified by LC–MS (details in Supplemental
methods) and the C60-K66CM peptide collected was used as a stock
solution (the concentration of the peptide was not determined).
Esterification of the C60-K66CM peptide was carried out by treatment
of 10
by 68
incubated at room temperature for 16 h, at which point 20 l
l
l
L C60-K66CM with 2
L of either water (control), 95% ethanol, or n-propanol and
L of
lL of concentrated sulfuric acid followed
9
96 Photodiode Array Detector with a SymmetryPrep™ C18
m 100 Å 19 ꢂ 150 mm) column (Waters) using a linear AB
gradient from 5% to 60% B for 12.5 min and 60% to 100% B for
.5 min and a flow rate of 10 mL/min (where Solvent A was 0.1%
(
7
l
1 M formic acid was added to the mixture and the samples were
analyzed by LC–MS (details in Supplemental methods).
2
aqueous TFA and Solvent B was 0.1% TFA in acetonitrile). The major
peak at 9 min was collected and the product concentrated until
2.5. Activity assays
ꢁ
2 mL remained. This was flash frozen in an acetone/dry ice bath
2.5.1. Inactivation and competitive inhibition of GLO1 by 4BAB
Inactivation of GLO1 by 4BAB was carried out at concentrations
of 4.7, 14.4, 31.0 and 80.6 lM in 75 mM HEPES, pH 7.3 and were
for 15 min then lyophilized to dryness. Clear/white crystals
remained with a percent yield of 20%. The 1H NMR was acquired
on a ECX 400 MHz NMR (JEOL, Peabody, MA, USA) and MS analysis
incubated with GLO1 at 25 °C ranging from 5 to 180 s after which
point GLO1 activity was measured as previously described.43
Enzymatic activities are reported as the initial rate of the reaction
measured by the change in OD240/min using a DU640 spectrometer
(Beckman Coulter, Brea, CA). The concentration of thiohemiacetal
was performed on a Apex IV Qh Ultra FT-ICR (Bruker Daltonics,
1
Fremont, CA, USA). Both the H NMR (d 1.76–1.80 (m, –O–CH
2
CH
2
–);
d 2.19–2.28 (m, Glu-C
.99 (q, Cys-C );
q, –S–CH –C(O)); d 4.02 (s, –OC(O)–CH
.05–4.09 (t, Glu-C H); d 4.21–4.27 (m, C(O)O–CH
.59–4.63 (q, Cys-C H) and MS analysis are consistent with the
b
H
2
); d 2.56–2.63 (m, Glu-C
Y
H
2
); d 2.93–
2
a
H
a
d
3.12–3.17 (q, Cys-C );
a
H
b
d
3.41–3.49
4
4
(
2
2
–Br), d 4.05 (s, Gly-CH
2
);
substrate used in the reaction was 0.18 mM and the concentra-
tion of free GSH was fixed at 0.20 mM for all assay solutions. Sub-
strate mixtures were allowed to equilibrate for 30 min in 50 mM
HEPES, pH 7.3 at 25 °C in a 1 cm path length cuvette with a final
4
4
a
2
–CH –); d
2
a
structure and theoretical mass of 4BAB (Supplemental Figs. S2–S4,
respectively).
total volume of 1 mL, after which point 100
tion solution was added to the cuvette to measure activity. After
the addition of enzyme, the initial rate ( ) of the enzyme catalyzed
reaction was obtained from the increase in absorbance at 240 nm
over one minute due to the formation of S- -lactoylglutathione
lL of enzyme inactiva-
2
.3. Purification of human GLO1 enzyme
v
0
Human GLO1 clone41 in a pKK223-3 expression vector was
D
ꢀ
1
ꢀ1 43
obtained as a gift from Prof. Bengt Mannervik at the University of
Uppsala, Sweden. The plasmid was introduced into Escherichia coli
JM103 cells as previously described using electroporation (2.5 kv,
(e
240 = 2860 M cm ).
ꢀ1
ꢀ1
ꢀ1
ðdA240=dtÞ min ꢂ 1000
lL
v0
ð
l
mol min Þ ¼
ꢀ
1
4
2
e
2
40
M
cm ꢂ 1:0 cm
2
5
l
F, and 400
X
). Cells were plated on Luria–Bertani (LB) aga-
g/mL of ampicillin to screen
rose plates supplemented with 100
l
The maximum velocity (Vmax) was subsequently calculated
for colonies that had been transfected with the plasmid. To ensure
that the positively transfected cells contained the correct GLO1
coding sequence, the plasmid of a selected single colony was
extracted using the Spin Miniprep Kit (Qiagen, Hilden, Germany)
from the Michaelis–Menten equation using K
m
= 0.18 mM for
n
is the rate at
4
4
GLO1. Plots of ln(
t = n and
Fig. S6a) were constructed to determine the extent of inactivation.
v
n
/v
0
) versus time (where
v
0
v is rate at t = 0, Fig. 2A) and Abs240 versus time
(
0
and sequenced with a primer having sequence 5 -GCAC TACA TTAA
For 4BAB competitive inhibition studies, GLO1 activity was mea-
sured as described above at varying substrate concentrations
0
GGTT GCCA TTTT GTTA GG-3 . The recombinant enzyme was over-
ꢀ
1
ꢀ1
expressed, purified, and quantified
(
e
280 = 1.67 mL mg cm
)
(
0.10–0.70 mM) in the presence of 0.00–157.6 lM 4BAB.
41
according to published procedures. Briefly, a single colony was
cultured in LB broth at 37 °C until a cell density of reached
2.5.2. Protection of GLO1 against 4BAB inactivation
A
600 ꢃ 0.3; the expression of GLO1 was induced for 12 h at 30 °C
GLO1 was incubated in 75 mM HEPES pH 7.3 at 25 °C with
by the addition of 0.2 mM isopropyl-b- -thiogalactopyranoside
D
13.1
at concentrations of 0.00
was measured and plots of ln( n
l
M 4BAB in the presence of the competitive inhibitor CHG
M, 0.108 M and 0.324 M. The activity
v0) versus time (Fig. 2D) and
and 1.0 mM zinc chloride. The bacteria were lysed using a French
Press, and the GLO1 protein was purified from the supernatant
using S-hexylglutathione agarose affinity chromatography as pre-
l
l
l
v
/
Abs240 versus time (Fig. S6b) were constructed as described above
to determine the extent of inactivation.
4
1
viously described
followed by gel-filtration chromatography
(
Sephadex G-50). The mass and purity of the enzyme were
assessed using both SDS and native polyacrylamide gel electropho-
resis. SDS–PAGE (4–12% Bis–Tris) resolved two bands with molec-
ular weights of approximately 21 and 42 kDa, which are consistent
with the amino acid deduced monomer (20,647 Da) and dimer
2.5.3. Kinetic analysis of GLO1:4BAB complex
A solution of GLO1:4BAB was made by mixing 100
2.3 mg/mL GLO1 stock with 500 L 100 mM HEPES pH 7.4,
300 L H O and 100 L 13.45 mM 4BAB (final concentrations:
lL of
l
l
2
l
(
(
41,294 Da) masses of GLO1, respectively (Fig. S5a). Native PAGE
4–16% Bis–Tris) showed one band consistent with the MW for
50 mM HEPES pH 7.4, 0.3 mg/mL GLO1, and 0.672 mM 4BAB).
The activity of the solution was monitored until there was no
change in activity over time. After 5 min, the activity stabilized
and the reaction was treated with 2 mL of H O and washed using
2
a Millipore Centriplus Filter (MW cutoff of 10,000 Da) at 3000 g
the GLO1 dimer (41,294 Da) confirming that only the dimer is
present under native conditions (Fig. S5b). ESI-MS analysis of the
purified protein showed multiply charged species that were
consistent with a MW of 20,647 Da for the enzyme monomer.
for 45 min at 4 °C to remove any unreacted 4BAB, with the
Please cite this article in press as: Holewinski, R. J.; Creighton, D. J. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.055

4
R. J. Holewinski, D. J. Creighton / Bioorg. Med. Chem. xxx (2014) xxx–xxx
Figure 2. Kinetic characterization of 4BAB on GLO1. (A) GLO1 inactivation by 4BAB expressed as % remaining activity versus time. (B) Michaelis–Menten kinetics for GLO1
control and GLO1:4BAB solutions. Kinetic constants are given in Table 1. (C) Dixon plot showing competitive inhibition of GLO1 by 4BAB. (D) Protection against inactivation of
GLO1 by 4BAB using the strong competitive inhibitor CHG (K = 46 nM).
i
subsequent addition of 2 mL of H
conditions. A control solution (GLO1 control) was prepared in the
same manner, substituting 100 L of water for 4BAB and enzyme
concentrations were determined by measuring the absorbance at
2
O and washed under the same
29 h. The digestion was stopped by addition of 10
lL 50% TFA
and 25 L TCEP and flash freezing the solution in an acetone/
l
l
dry ice bath. Tryptic peptides were analyzed by LC–MS/MS
(details in Supplemental methods).
ꢀ1
ꢀ1
41
2
80 nm using
e280 1.67 mL mg cm
for GLO1. The starting
g/mL for the GLO1
enzyme concentrations were 1.24 and 1.20
l
3. Results
control and GLO1:4BAB solution, respectively. Activities at various
substrate concentrations (0.05–0.8 mM thiohemiacetal) were
3.1. Synthesis of BBE and 4BAB
measured in the presence of 50 mM HEPES pH 7.4 using 100
of the working enzyme solution.
lL
1
Synthesis of BBE and 4BAB was confirmed by 400 MHz H NMR
(Figs. S1 and S2) and 4BAB was confirmed by high resolution MS
2
.6. Sample preparation and LC–MS analysis of intact and
analysis (Figs. S3 and S4). Proton NMR peaks for 4BAB correspond-
ing to those of the glutathione moiety were assigned as previously
trypsin digested GLO1:4BAB complex
4
5
reported. The remaining peaks corresponding to 4BAB were
1
The GLO1 control and GLO1:4BAB solutions prepared for the
kinetic analysis above were analyzed by LC–MS with and with-
assigned based on the H NMR of BBE. ESI-MS was consistent
with the singly charged monoisotopic mass of 4BAB (theoretical
monoisotopic singly charged m/z = 558.07515, observed m/z =
558.09332) and the spectrum showed the splitting due to
out tryptic digestion. For each trypsin digestion, 10
control and GLO1:4BAB solutions were treated with 90
00 mM HEPES pH 7.3 followed by 300 L of digestion solution
containing 6 M urea, 2 mM DTT, 50 mM NH HCO pH 8.05 and
incubated at 95 °C for 30 min. The solution was cooled to room
temperature and treated with 300 L of dilution buffer contain-
ing 50 mM NH HCO pH 8.05, 1 mM CaCl
trypsin and incubated at 37 °C for 4 h. Three additional 4
lL of GLO1
l
L of
7
9
81
1
l
the two isotopes (Br and Br ) of bromine (Fig. S4), which are
approximately equally abundant in nature.
4
3
l
3.2. Inactivation/covalent modification of GLO1 by 4BAB
4
3
2
and 4
l
L of 0.5
l
g/lL
L
l
Human GLO1 was inactivated by 4BAB (Figs. 2A and S6a),
although activity of GLO1 never reached zero. MS analysis of the
portions of trypsin were added over a total incubation time of
Please cite this article in press as: Holewinski, R. J.; Creighton, D. J. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.055

R. J. Holewinski, D. J. Creighton / Bioorg. Med. Chem. xxx (2014) xxx–xxx
5
GLO1:4BAB complex (Fig. S6) showed two peaks, the first peak
consistent with the mass of the GLO1 monomer modified by
for the GLO1:4BAB complex (singly charged m/z = 911.4, Fig. 3C
and D). Tandem MS analysis and carboxylic acid side chain
esterification of the C60-K66 peptide confirmed Cys60 to be the site
of modification (Figs. 4 and 5).
4
BAB (21,125 Da) and the second peak consistent with the unmod-
ified monomer mass (20,647 Da). A comparison of the kinetics of
substrate turnover and binding by 4BAB modified and unmodified
GLO1 was carried out (Fig. 2B, Table 1). Data for the competitive
inhibition of GLO1 by 4BAB is illustrated in Fig. 2C. The effect on
4. Discussion
4
BAB inactivation of GLO1 by the addition of the transition state
Although the synthesis of 4BAB is straightforward, the final
yields of 4BAB only ranged from 9% to 25%. The reason for the
low yields is not entirely clear, but possible sources of product loss
could be the breakdown of product during washes, formation of
the di-GSH substituted diester, or loss of product during purifica-
tion. The starting dibromo-diester is in 5-fold excess over GSH,
which should prevent, or at least limit, the formation of the
di-GSH substituted diester. However, H NMR of the HPLC peak
preceding the product peak (data not shown) shows that this side
product is formed in the reaction mixture, which in turn decreases
analogue CHG (K
i
= 46 nM)34 is indicated in Figs. 2D and S6b. The
interaction of 4BAB was confirmed to be covalent by LC–MS anal-
ysis of the GLO1:4BAB complex (Fig. S7), as there is a peak present
with m/z values consistent with the GLO1 monomer modified by
4
BAB (minus the bromine, which has been displaced) (Table S1).
The modification was localized to the tryptic peptide comprising
the amino acid sequence C60-K66 as evident by the reduction in
ion intensity for the C60-K66 peptide (singly charged m/z = 853.5,
Fig. 3A and B) and appearance of a new ion species was observed
1
Table 1
Kinetic constants for GLO1 control and the GLO1:4BAB complex. Values are given as the mean ± S.D
ꢀ
1
ꢀ1
ꢀ1
ꢀ1
l )
M -1
Enzyme
Specific activity (
l
mol min mg
)
k
cat (s
)
K
m
(l
M)
kcat/K
m
(s
GLO1
GLO1:4BAB
1580 ± 52
516 ± 22
1088 ± 36
355 ± 15
182 ± 18
221 ± 26
5.98 ± 0.62
1.61 ± 0.21
Figure 3. MS analysis of 4BAB interaction with GLO1. Extracted ion currents (XICs) for ion 853.5 ± 0.5 Da for (A) GLO1 control and (B) GLO1:4BAB trypsin digests as well as
XICs for 911.4 ± 0.5 Da for (C) GLO1 control and (D) GLO1:4BAB trypsin digests.
Please cite this article in press as: Holewinski, R. J.; Creighton, D. J. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.055

6
R. J. Holewinski, D. J. Creighton / Bioorg. Med. Chem. xxx (2014) xxx–xxx
Figure 4. Collision induced dissociation of C60-K66 and C60-K66CM peptides. (A) Fragmentation of parent ion 853.4 m/z eluting at 53 min after trypsin digestion of GLO1,
consistent with peptide C60-K66. The parent ion and all fragment ions are singly charged. (B) Fragmentation of parent ion 911.4 m/z eluting at 57 min from trypsin digestion of
GLO1:4BAB. The 911.4 m/z is consistent with modification of C60-K66 by a carboxymethyl group and the asterisk indicates carboxymethyl modification. The parent ion and all
fragment ions are singly charged and the fragmentation pattern limits the site of modification to either Cys60 or Asp61.
Figure 5. LC/MS analysis of the C60-K66CM control, ethanol, and propanol esterification reactions. The TIC chromatogram for the control (A) shows a single peak with m/z
2
+
consistent with doubly charged C60-K66CM peptide (D), [C60-K66CM+2H] = 456.3 m/z. TIC chromatograms for the ethanol (B) and propanol (C) esterification of C60-K66CM.
The first peak in B and C corresponds to the mono-esterified C60-K66CM peptide, the second and third peaks correspond to variants of the di-esterified C60-K66CM peptide, and
⁄
#
the last peak corresponds to the tri-ethyl esterified C60-K66CM peptide (E, C60-K66CM ) and tri-propyl esterified C60-K66CM peptide (F, C60-K66CM ). Full chromatograms and
MS spectra can be found in the supplemental material.
the overall yield of desired product. Additionally, it is possible that
the leaving group is hydrolyzed during the wash cycles. The pH of
the GSH solution used in the reaction was ꢁ9.5 with the reaction
stopped by the addition of formic acid until the pH is 3.5. This
should prevent the hydrolysis of the leaving group, although the
low pH could make the ester bonds susceptible to hydrolysis. There
is also the possibility of other side reactions that have not been
accounted for here.
active site and orients the bromine near Cys60 for nucleophilic
attack, which should completely abolished enzyme activity. This
led to the question of how 4BAB could be covalently modifying
GLO1 but result in partial enzyme inactivation. One explanation
was that 4BAB could possibly be binding to GLO1 somewhere
other than the enzyme active site, which results in partial loss
of enzyme activity. However, this does not seem likely, since it
was shown that the transition state analogue CHG protects
GLO1 from inactivation by 4BAB (Fig. 2D), which suggests that
inactivation is an active site directed process. Therefore, another
molecular phenomenon must be causing the observed partial
inactivation.
The kinetic data obtained from incubation of 4BAB with GLO1
show that 4BAB does not completely inactivate the enzyme
(Figs. 2A and S6a). At low concentrations (4.7 and 14.4
lM),
4
BAB seemed to follow biphasic kinetics (Fig. 2A); even though
regardless of the concentration 4BAB, there was a plateau at
approximately the same level of maximal activity. This was quite
unexpected as we had proposed that 4BAB enters the enzyme
Another possibility is that there is active GLO1 monomer pres-
ent in the assay solution, and it is this species that inactivated by
4BAB leading to the observed activity decrease. However this most
Please cite this article in press as: Holewinski, R. J.; Creighton, D. J. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.055

R. J. Holewinski, D. J. Creighton / Bioorg. Med. Chem. xxx (2014) xxx–xxx
7
likely is not the case since the native PAGE shows that GLO1
Fig. S5b) is present exclusively as the dimer. Furthermore, there
present in the chromatograms for the GLO1:4BAB trypsin solution,
corresponding to a modified C60-K66 peptide. Modification of the
(
are no reports to our knowledge of monomeric human GLO1
retaining enzyme activity. This led to the hypothesis that only
one active site of the GLO1 homodimer is being modified by
C60-K66 peptide by 4BAB should produce a singly charged m/z value
of 1330.5. However, no new peaks corresponding to the singly or
multiply charged ions consistent with this modification were
present in the GLO1:4BAB trypsin digests.
4
BAB. To test this hypothesis, GLO1 treated with 4BAB was ana-
lyzed by LC–MS and two species were observed in the reaction
mixture; one that corresponded to the unmodified GLO1 monomer
and one that corresponded to GLO1 monomer with 4BAB (minus
the bromine) attached. Furthermore, the two species are present
in nearly equal quantities, although the peaks were not sufficiently
resolved to integrate or quantify (Fig. S7). If there was only one
active site of GLO1 modified, then it would be expected to maintain
Unexpectedly, the intact inactivator was not observed on the
60-K66 peptide, and therefore we set out to explore the possible
C
reasons for the absence of the modified C60-K66 peptide in the chro-
matograms. One possible explanation could be a missed cleavage
at Lys59 resulting from increased steric hindrance around this
residue, possibly making it inaccessible to trypsin, but this does
not seem to be the case since there is no ion present in the
5
0% of its activity but only 33% activity remains. We proposed that
chromatographs consistent with modification of the V51-K66
modification of one active site of GLO1 could cause a conforma-
tional change affecting the other active site present in the dimeric
form. The kinetic constants were measured for GLO1 control and
peptide. Another possible explanation could be that the ester
functions of the bound 4BAB are hydrolyzed during trypsin diges-
tion, as trypsin has been shown to possess esterase activity
for the GLO1:4BAB complex (Fig. 2B, Table 1) and the kcat and K
m
towards benzoyl-
L
-arginine ethyl ester (BAEE) and other
L-arginine
4
8,49
values obtained for GLO1 are in good agreement with values previ-
ously reported for the enzyme.44 The GLO1:4BAB enzyme complex
showed reduced catalytic efficiency such that there was a 2/3 loss
of catalytic turnover (decreased kcat) and about 1/5 decrease in
substrate binding (increased K ) for the GLO1:4BAB enzyme com-
m
plex compared to the wild-type enzyme. This is consistent with the
observation that GLO1 retains about 33% activity in the presence of
esters.
We predicted the possible species (Table S2) that would
arise from hydrolysis of these esters, assuming that C60-K66 is the
peptide modified, and checked the chromatograms for the
existence of these ions. Monitoring the extracted ion currents of
these ions showed the presence of an additional peak in the
GLO1:4BAB tryptic digest with an elution time of 57 min that
was not present in the control mixture (Fig. 3C), and this new peak
was due to the presence of a singly charged ion with m/z of 911.4
(Fig. 3D). This mass is consistent with the C60-K66 peptide plus the
addition of a carboxymethyl group (58 mass units), one of the
predicted species resulting from the hydrolysis of the ester bonds
of 4BAB (C60-K66CM, Table S2). LC–MS/MS analysis confirmed that
the modification was localized to either Cys60 or Asp61, both of
which could act as nucleophiles to displace the bromine. First,
4
BAB (Fig. 2A). A scenario could be envisioned in which one mole-
cule of 4BAB binds covalently to a single active site resulting in
complete inactivation simultaneously causing a conformational
change around the opening to the hydrophobic pocket of second
active site that results in the reduced kinetic parameters observed.
This conformational change could presumably prevent a second
molecule of 4BAB from orienting the bromine group near Cys60
of the second active site for attack. The inaccessibility to the hydro-
phobic pocket should not have that much of an effect on the bind-
ing of the thiohemiacetal substrate (about 20% decrease observed),
since the methyl group of substrate is much smaller than the tail
region of 4BAB. The proposal that modification at one active site
affects the binding at the second active site seems contradictory
the mass increase of 58 units of the b
the modification must be somewhere on peptide fragment
60-D-F-P63. Second, the fact that the y fragment is unchanged
4
ion in Fig. 4B indicates that
C
5
in Figure 4B means that residues F62-P-I-M-K66 are excluded as
possibilities. This leaves only residues Cys60 and Asp61 as the
possible sites of modification. Since low mass fragments in the
MS2 spectra were not detected, localization of the modification
site had to be achieved through alternative methods.
3
6
to the results reported for bivalent transition state analogues,
which have a lower K than the transition state molecules by them-
i
selves. However, these bivalent transition state analogues bind to
the active site in the same manner as a single transition state
molecule, which is by coordination of the N-hydroxycarbamoyl
moiety of the molecule to the active site zinc ion. There has yet
to be evidence that indicates that GLO1 has cooperative
The presence of a carboxymethyl group on either Cys60 or
Asp61 would yield two structurally different peptides with similar
but different chemical properties (referred to as C60-K66CMCys if
modification is at Cys60 and C60-K66CMAsp if modification is at
Asp61, Fig. S8). Carboxymethylation of Cys60 results in a peptide
with three free carboxylic acid functions and carboxymethylation
of Asp61 results in only two free carboxylic acid functions.
Therefore, treatment of the purified C60-K66CM peptide with
3
4,41
binding with regards to substrate or transition state analogues.
Allosteric coupling has been reported for the two active sites of the
monomeric Plasmodium falciparum GLO146 and was proposed for
the two active sites of monomeric yeast GLO147 but no reports of
allosteric binding have been reported for the dimeric human
enzyme.
2 4
ethanol and propanol in the presence of concentrated H SO
should result in esterification of the free carboxylic acid residues
on the peptide through a traditional Fisher esterification mecha-
5
0
Given that inactivation of GLO1 by 4BAB is an active site
directed mechanism, as the transition state analogue CHG protects
GLO1 against inactivation by 4BAB (Figs. 2D and S6b), we next set
out to determine the amino acid that is covalently modified by
nism. The m/z values of the mono- and di-esterified C60-K66CM
peptide would be the same if either Cys60 or Asp61 were carbo-
xymethylated (Table S3). The m/z values for the peaks eluting at
87 and 95 min (Fig. 5, Table S3) for the ethanol and propanol
esterification, respectively, are consistent with the tri-esterified
4
BAB. From the X-ray crystallographic structures1,2 Cys60 is
located in the hydrophobic binding pocket and is the most likely
candidate to act as the nucleophile to displace the bromine on
C60-K66CM peptide, which could only arise if the carboxymethyla-
tion occurs on Cys60. The MS/MS data presented in Figure 4 and
the esterification data presented in Figure 5 confirms that Cys60
is the site of covalent modification between GLO1 and 4BAB.
4
BAB. To test this hypothesis, the GLO1:4BAB complex was
digested with trypsin and analyzed by LC–MS/MS and based on
the in silico tryptic peptides predicted for GLO1, Cys60 would lie
on a 7 amino acid long peptide (C60-D-F-P-I-M-K66, C60-K66).
Complete disappearance of the 853.4 m/z ion (corresponding to
the singly charged C60-K66 peptide) was not observed, consistent
with incomplete inactivation of the enzyme, leaving a fraction of
the Cys60 residues unmodified. There should also be a new peak
5. Conclusion
This is the first reported inhibitor, 4BAB, that covalently
modifies an amino acid residue near the active site of GLO1. This
inhibitor specifically modifies Cys60 located in the hydrophobic
Please cite this article in press as: Holewinski, R. J.; Creighton, D. J. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.055

8
R. J. Holewinski, D. J. Creighton / Bioorg. Med. Chem. xxx (2014) xxx–xxx
binding pocket adjacent to the GLO1 active site; however, this
modification results in complete inactivation of only one active
site, although leading to impaired catalytic efficiency at the second
active site. We acknowledge that our inhibitor 4BAB is a lead
molecule and as it is will have limited use pharmacologically.
Rather our goal was a proof of principle that GLO1 could be
inhibited covalently ay Cys60 within the enzyme active site.
While we have not tested the specificity of 4BAB to GLO1 this
molecule may serve as a template for the development of new
GLO1 inhibitors that may combine this strategy with ones already
reported for high affinity GLO1 inhibitors, potentially improving
potency and specificity.
17. Kuhla, B.; Boeck, K.; Schmidt, A.; Ogunlade, V.; Arendt, T.; Munch, G., et al
Neurobiol. Aging 2007, 28, 29.
18. Loos, M.; van der Sluis, S.; Bochdanovits, Z.; van Zutphen, I. J.; Pattij, T.; Stiedl,
O., et al Genes Brain Behav. 2009, 8, 817.
19. Distler, M. G.; Plant, L. D.; Sokoloff, G.; Hawk, A. J.; Aneas, I.; Wuenschell, G. E.,
et al J. Clin. Invest. 2012, 122, 2306.
20. Benton, C. S.; Miller, B. H.; Skwerer, S.; Suzuki, O.; Schultz, L. E.; Cameron, M. D.,
et al Psychopharmacology 2011, 221, 297.
21. Distler, M. G.; Gorfinkle, N.; Papale, L. A.; Wuenschell, G. E.; Termini, J.; Escayg,
A., et al Epilepsia 2013, 54, 649.
22. Zhang, D.; Tai, L. K.; Wong, L. L.; Chiu, L. L.; Sethi, S. K.; Koay, E. S. Mol. Cell.
Proteomics 2005, 4, 1686.
23. Rulli, A.; Carli, L.; Romani, R.; Baroni, T.; Giovannini, E.; Rosi, G., et al Breast
Cancer Res. Treat. 2001, 66, 67.
2
2
4. Fonseca-Sanchez, M. A.; Rodriguez Cuevas, S.; Mendoza-Hernandez, G.;
Bautista-Pina, V.; Arechaga Ocampo, E.; Hidalgo Miranda, A.; Quintanar
Jurado, V.; Marchat, L. A.; Alvarez-Sanchez, E.; Perez Plasencia, C.; Lopez-
Camarillo, C. Int. J. Oncol. 2012, 41, 670.
5. Wang, Y.; Kuramitsu, Y.; Ueno, T.; Suzuki, N.; Yoshino, S.; Iizuka, N., et al
Anticancer Res. 2012, 32, 3219.
Acknowledgments
We would like to thank Dr. Guo-Zhang Zhu for his help in
purification of recombinant human GLO1. We would also like to
thank Dr. James Fishbein at UMBC for his assistance in guiding
some of the final experimental approaches to this manuscript after
the passing of Dr. Donald Creighton.
26. Bair, W. B., 3rd; Cabello, C. M.; Uchida, K.; Bause, A. S.; Wondrak, G. T.
Melanoma Res. 2010, 20, 85.
2
2
2
7. Antognelli, C.; Mezzasoma, L.; Fettucciari, K.; Talesa, V. N. Int. J. Biochem. Cell
Biol. 2013, 45, 836.
8. Antognelli, C.; Mezzasoma, L.; Fettucciari, K.; Mearini, E.; Talesa, V. N. Prostate
2012, 73, 121.
9. Antognelli, C.; Baldracchini, F.; Talesa, V. N.; Costantini, E.; Zucchi, A.; Mearini,
E. Cancer J. 2006, 12, 222.
3
3
3
0. Vince, R.; Daluge, S. J. Med. Chem. 1971, 14, 35.
1. Vince, R.; Daluge, S.; Wadd, W. B. J. Med. Chem. 1971, 14, 402.
2. Thornalley, P. J.; Ladan, M. J.; Ridgway, S. J.; Kang, Y. J. Med. Chem. 1996, 39,
3409.
3. Sakamoto, H.; Mashima, T.; Sato, S.; Hashimoto, Y.; Yamori, T.; Tsuruo, T. Clin.
Cancer Res. 2001, 7, 2513.
4. Murthy, N. S.; Bakeris, T.; Kavarana, M. J.; Hamilton, D. S.; Lan, Y.; Creighton, D.
J. J. Med. Chem. 1994, 37, 2161.
5. Sharkey, E. M.; O’Neill, H. B.; Kavarana, M. J.; Wang, H.; Creighton, D. J.; Sentz,
D. L., et al Cancer Chemother. Pharmacol. 2000, 46, 156.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2014.04.055.
3
3
3
References and notes
1
2
.
.
Cameron, A. D.; Ridderstrom, M.; Olin, B.; Kavarana, M. J.; Creighton, D. J.;
Mannervik, B. Biochemistry 1999, 38, 13480.
Cameron, A. D.; Olin, B.; Ridderstrom, M.; Mannervik, B.; Jones, T. A. EMBO J.
36. Zheng, Z. B.; Creighton, D. J. Org. Lett. 2003, 5, 4855.
37. More, S. S.; Vince, R. Bioorg. Med. Chem. Lett. 2006, 16, 6039.
38. More, S. S.; Vince, R. J. Med. Chem. 2009, 52, 4650.
39. Vince, R.; Brownell, J.; Akella, L. B. Bioorg. Med. Chem. Lett. 1999, 9, 853.
40. Witschi, A.; Reddy, S.; Stofer, B.; Lauterburg, B. H. Eur. J. Clin. Pharmacol. 1992,
43, 667.
41. Ridderstrom, M.; Mannervik, B. Biochem. J. 1996, 314, 463.
42. Sambrook, J.; Fritsch, E. M.; Maniatis, T. Molecular Cloning: A Laboratory Manual;
Cold Spring Harbor Laboratory Press: New York, 1989.
43. Vander Jagt, D. L.; Han, L. P.; Lehman, C. H. Biochemistry 1972, 11, 3735.
44. Hamilton, D. S.; Creighton, D. J. Biochim. Biophys. Acta 1992, 1159, 203.
45. Rabenstein, D. L.; Keire, D. A. In Glutathione: Chemical, Biochemical, and
Medicinal Aspects; Dolphin, D., Avramovic, O., Poulson, R., Eds.; J. Wiley &
Sons: New York, 1988; p 67.
1
997, 16, 3386.
3
4
5
6
7
8
9
.
.
.
.
.
.
.
Cameron, A. D.; Ridderstrom, M.; Olin, B.; Mannervik, B. Structure 1999, 7, 1067.
Racker, E. J. Biol. Chem. 1951, 190, 685.
Richard, J. P. Biochemistry 1991, 30, 4581.
Phillips, S. A.; Thornalley, P. J. Eur. J. Biochem. 1993, 212, 101.
Thornalley, P. J.; Langborg, A.; Minhas, H. S. Biochem. J. 1999, 344, 109.
Rabbani, N.; Thornalley, P. J. Ann. N.Y. Acad. Sci. 2008, 1126, 124.
Lo, T. W.; Westwood, M. E.; McLellan, A. C.; Selwood, T.; Thornalley, P. J. J. Biol.
Chem. 1994, 269, 32299.
1
1
1
1
0. Bucala, R. Exp. Physiol. 1997, 82, 327.
1. Thornalley, P. J. Biochem. Soc. Trans. 2003, 31, 1372.
2. Xue, M.; Rabbani, N.; Thornalley, P. J. Semin. Cell Dev. Biol. 2011, 22, 293.
3. Rabbani, N.; Chittari, M. V.; Bodmer, C. W.; Zehnder, D.; Ceriello, A.; Thornalley,
P. J. Diabetes 2010, 59, 1038.
46. Deponte, M.; Sturm, N.; Mittler, S.; Harner, M.; Mack, H.; Becker, K. J. Biol. Chem.
2007, 282, 28419.
1
1
1
4. Berlanga, J.; Cibrian, D.; Guillen, I.; Freyre, F.; Alba, J. S.; Lopez-Saura, P., et al
Clin. Sci. (London) 2005, 109, 83.
5. Ahmed, N.; Babaei-Jadidi, R.; Howell, S. K.; Thornalley, P. J.; Beisswenger, P. J.
Diabetes Care 2005, 28, 2465.
6. Luth, H. J.; Ogunlade, V.; Kuhla, B.; Kientsch-Engel, R.; Stahl, P.; Webster, J., et al
Cereb. Cortex 2005, 15, 211.
47. Frickel, E. M.; Jemth, P.; Widersten, M.; Mannervik, B. J. Biol. Chem. 2001, 276,
1845.
48. Schwert, G. W.; Eisenberg, M. A. J. Biol. Chem. 1949, 179, 665.
49. Anderson, R. A., Jr.; Beyler, S. A.; Mack, S. R.; Zaneveld, L. J. Biochem. J. 1981, 199,
307.
50. Fischer, H. E.; Speier, A. Chem. Ber. 1895, 28, 3252.
Please cite this article in press as: Holewinski, R. J.; Creighton, D. J. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.055